All this H1B stuff, it's nonsense.
The tech companies are hiring subpar developers who can barely keep pace with the dimmest bulb among the whites. The difference between the whites and the indians / chinese is the non-whites will lie to each other and their bosses about the state of the project. Hence, you get stuff like Zuckerberg unable to answer a phone call with his new glasses project. He was using a product based on lies.
I left the tech industry because it was dying. There was nothing new to write, nothing new to create, because all of the companies had been lying to the world about their tech and the lies were becoming more and more apparent as time went on. The internet doesn't solve many problems. Most people don't need it. At best, it's just a faster USPS or a more configurable telephone. Most of what people use it for can be handled by a telegraph service or fax machines.
The straw that broke the camel's back was when I discovered the huge amount of deception surrounding "AI". The models they were using were never going to work. None of the research suggested it would ever work. Even for simple things like character recognition and whatnot it was a terrible disaster, consuming way too many compute resources for otherwise simple tasks. And at best you'd get something like 85% accuracy. The amount of data and compute cycles needed to get to 90% was astronomical, and it was all exponential from there on. Somehow, babies were getting it right 95% of the time and they had a fraction of the compute power of what we were using to get it right 80% of the time.
Combine that with the fact that too many people were outright LYING about their results or their methodology. Remember the guy who started an AI company that was simply having underpaid indians do the work for them? Yes, the "mechanical turk" is still the best AI there is, and it always will be. Hire a guy, pay him a few dollars, and you'll get far more bang for the buck than you will trying to make a machine do the simplest task.
Anyway, the tech industry is what's holding our entire economy together. Most companies post modest profits and growth, which is perfectly fine if the inflation rate is sane or the economy is steadily growing. But it's not good enough to support the insane amount of investment we have in our economy. People are literally betting that we'll grow by 10% or more each year, and then being lied to about it. It's all a scam, and the heart of the scam is tech companies.
One day, tech companies will be valuated at their true value, and most people are going to lose all their wealth because of it. H1Bs are how tech companies are trying to maintain their value despite not having anything of value at all. They want to fill seats and deliver useless products built on lies just to keep the scam going for another few years.
One day we're going to wake up and realize that all of these foreign workers were just scamming us on an epic scale. They all have to go back. The sooner they leave the better it is for them. It's like quitting a company before the layoffs start. It's better to do that because you have a chance at finding a new job. Once the layoffs start, you're never going to find a job again.
They can take Facebook, Microsoft, Amazon, and all the other tech companies with them. I don't care. They don't provide any economic value to our country anymore anyway.
The tech companies are hiring subpar developers who can barely keep pace with the dimmest bulb among the whites. The difference between the whites and the indians / chinese is the non-whites will lie to each other and their bosses about the state of the project. Hence, you get stuff like Zuckerberg unable to answer a phone call with his new glasses project. He was using a product based on lies.
I left the tech industry because it was dying. There was nothing new to write, nothing new to create, because all of the companies had been lying to the world about their tech and the lies were becoming more and more apparent as time went on. The internet doesn't solve many problems. Most people don't need it. At best, it's just a faster USPS or a more configurable telephone. Most of what people use it for can be handled by a telegraph service or fax machines.
The straw that broke the camel's back was when I discovered the huge amount of deception surrounding "AI". The models they were using were never going to work. None of the research suggested it would ever work. Even for simple things like character recognition and whatnot it was a terrible disaster, consuming way too many compute resources for otherwise simple tasks. And at best you'd get something like 85% accuracy. The amount of data and compute cycles needed to get to 90% was astronomical, and it was all exponential from there on. Somehow, babies were getting it right 95% of the time and they had a fraction of the compute power of what we were using to get it right 80% of the time.
Combine that with the fact that too many people were outright LYING about their results or their methodology. Remember the guy who started an AI company that was simply having underpaid indians do the work for them? Yes, the "mechanical turk" is still the best AI there is, and it always will be. Hire a guy, pay him a few dollars, and you'll get far more bang for the buck than you will trying to make a machine do the simplest task.
Anyway, the tech industry is what's holding our entire economy together. Most companies post modest profits and growth, which is perfectly fine if the inflation rate is sane or the economy is steadily growing. But it's not good enough to support the insane amount of investment we have in our economy. People are literally betting that we'll grow by 10% or more each year, and then being lied to about it. It's all a scam, and the heart of the scam is tech companies.
One day, tech companies will be valuated at their true value, and most people are going to lose all their wealth because of it. H1Bs are how tech companies are trying to maintain their value despite not having anything of value at all. They want to fill seats and deliver useless products built on lies just to keep the scam going for another few years.
One day we're going to wake up and realize that all of these foreign workers were just scamming us on an epic scale. They all have to go back. The sooner they leave the better it is for them. It's like quitting a company before the layoffs start. It's better to do that because you have a chance at finding a new job. Once the layoffs start, you're never going to find a job again.
They can take Facebook, Microsoft, Amazon, and all the other tech companies with them. I don't care. They don't provide any economic value to our country anymore anyway.
I think something like this may already exist but is not very well marketed or popular to consumers.
Basically i want to be able to detect water leaks, remotely. I think something like this could even reduce insurance premiums and provide extra level of protection, maybe even get government support if we prove it reduces water usage or detects water leaks better.
Something like this likely already exists commercially for large water supply lines but isnt marketed to residential consumers
I also want to market counter top ice machines. I still know people who spend $5-$10 per day buying bags of ice from gas stations. I want a counter top machine that makes ice, maybe even bags it for you. Grab and go. Or maybe a unit about the size of a kitchen trash can. Refrigerator ice machines are notorious for failing even on high end premium consumer grade refrigerators. Remove the constantly failing component from refrigerator and market it as a premium, separate unit dedicated to dispensing and bagging ice. $$$ market it to the people who typically buy bags of ice. Maybe even market an ice machine that fits inside back seat cab area of a truck and runs off inverter or 12volt supply transformer. Can run off 1/4 water supply line or a refillable tank or we could even design it with built in dehumidifier that pulls water out of the air and turns it into dispensable ice making it both a dehumidifier and ice bagging machine in one.
I also think we should design and market wifi connected smoke detectors. Again, this is to monitor an area remotely. Being able to remotely detect smoke or water leaks. There's still a market for this internet of things stuff that has received hesitation and fear because people don't trust the chinese gadgets or the privacy of these devices. So the branding would have to emphasize the privacy of these devices or the ability to hook them to your own secure networks and eliminate third party servers or spyware.
I also think the internet ventured away from people making their own websites. Back in the 90s when teenagers got on the internet one of the first things everyone wanted to do was make their own websites. Yet no one does this anymore. They just get on social media which is corporate triple letter agency ghettos and prison yards. Let's get back to designing and hosting our own websites and decentralizing the internet to make it more interesting. The internet also needs better search engines like we used to have. Google is ruined and the internet is crawling with bots.
The pipe-monitoring idea is a good one!
1L/min - 30 L/min brass flow meters that fit residential pipes and hose fittings are ~$15 on amazon
This sort of job -- monitoring a colorful panel of sensors, programming the sensors and such -- will always be a job that needs humans to do it. AI can not do it.
The problem with refrigerator ice machines is that they try to pack so much in a tiny area. It's all designed to break at a certain point too. They want you coming back to buy another fridge every 5 years.
The issue with detectors of any sort is information overload. I've worked on software systems with millions of sensors and at any point in time, a certain number of them are in a false alert. No sensor is going to be 100% accurate. You're going to get a lot of false negatives and a whole lot of false positives. It's the false negatives that cause problems, so you're tempted to turn the sensitivity up. Now your whole alert panel is blinking. What will you do then, just ignore the alarms as "normal"?
It is impossible for any program to "learn" that a certain state of alarms is "normal" or not. But if you pay someone to sit in a seat and look at the blinking lights they can figure it out pretty quick. With enough feedback they'll tell you when it's time to buy stock or when to sell too. NASA was rumored to have a guy look at an immense amount of data scrolling past on the screen and raise his hand if something felt off. He was remarkably accurate for identifying critical conditions.
The best way to solve water leaks is to get eyes on the meter and get eyes on the water appliances. Have your people periodically inspect things. For instance, in my house, everyone knows they have to listen for the toilet water to stop running before they leave the bathroom. If the toilet water doesn't stop running, they have to fix it. Again, something that technology can't solve.
95% of the time everything works perfectly. Or 100% of the time 95% of the stuff is working. Either way, you need someone with eyes, ears, hands and a brain poking around if you want to get close to 100%. Technology will only get you close and make the user more effective at his tasks.